Intelligent design method for precision injection mold based on deformation field collaborative control

Abstract

The application discloses a kind of precision injection mold intelligent design methods based on deformation field collaborative regulation, including the digital twin of mould is constructed;Production data in injection molding process is collected, and is synchronously mapped to digital twin to carry out dynamic simulation, the deformation trend of key area of mould is predicted online, the influence factor and contribution degree leading to predicted deformation are analyzed reversely and traced, and trace analysis result is generated;Deformation trend and trace analysis are matched from collaborative regulation strategy library to collaborative adjustment scheme;The scheme is executed and the production data after execution is continuously collected and fed back to digital twin, the actual deformation of mould after regulation is compared with prediction target to determine regulation error, and the simulation model of digital twin and regulation strategy library are self-optimized.Life of mould is prolonged, and production cost is reduced.

Description

Technical Field

[0001] This invention relates to the field of injection mold technology, and in particular to an intelligent design method for precision injection molds based on the coordinated control of deformation fields. Background Technology

[0002] In the field of precision injection mold design and production control, existing technologies largely rely on traditional simulation models and experience-based control methods. Traditional simulation models often focus on single-physics field analysis, lacking a comprehensive consideration of the interactions of multiple physical fields such as heat, flow, and force during the injection molding process. This results in insufficient adaptability of the models to actual production scenarios, and significant deviations between simulation results and real-world conditions. Furthermore, existing technologies often treat production data acquisition and simulation analysis as separate processes, failing to dynamically track the deformation trends of the mold throughout the entire injection molding process. Deformation data can only be obtained through offline analysis or post-processing inspection, leading to predictive lag issues.

[0003] Furthermore, when performing deformation control, most existing technologies adopt a single-dimensional parameter adjustment strategy, lacking a systematic source analysis of deformation influencing factors, resulting in a strong degree of blindness in the formulation of control schemes. Due to the lack of an effective closed-loop feedback mechanism, the data after control cannot be fed back to the model optimization in a timely manner, and the adaptability of the simulation model and control strategy is difficult to improve with the iteration of production conditions. After multiple controls, there may still be large deformation errors, which cannot meet the stringent requirements of high-precision injection molded products for mold deformation control, ultimately leading to problems such as low product qualification rate, limited production efficiency, and shortened mold life. Summary of the Invention

[0004] To address at least one of the aforementioned technical problems, this invention provides an intelligent design method for precision injection molds based on the coordinated control of deformation fields.

[0005] In a first aspect, the present invention provides an intelligent design method for precision injection molds based on the coordinated control of deformation fields, the method comprising:

[0006] Based on the geometric model, material properties, sensor layout, and historical production data of the mold, a virtual mold model with multi-physics coupling simulation capabilities including heat-fluid-force is constructed as a digital twin.

[0007] Collect production data during the mold injection process, including the process parameters of the injection molding machine and the mold status data; synchronously map the production data to a digital twin for dynamic simulation, predict the deformation trend of key areas of the mold online, and obtain the trend prediction results;

[0008] Based on the trend prediction results, the influencing factors and their contributions that cause the prediction distortion are analyzed in reverse and traced back to generate source analysis results; based on the trend prediction results and source analysis results, a collaborative adjustment scheme is dynamically matched from a pre-set collaborative control strategy library;

[0009] The collaborative adjustment plan is implemented, and production data after implementation is continuously collected and fed back to the digital twin. The actual deformation of the mold after adjustment is compared with the predicted target to determine the adjustment error. The adjustment error is used to perform self-learning optimization of the simulation model of the digital twin and the adjustment strategy library.

[0010] Secondly, the present invention also provides an intelligent design system for precision injection molds based on the coordinated control of deformation fields, the system comprising:

[0011] The digital twin building block is used to construct a virtual mold model with multi-physics coupling simulation capabilities, including thermal-fluid-mechanical fields, based on the mold's geometric model, material properties, sensor layout, and historical production data, thus serving as a digital twin.

[0012] The deformation trend prediction unit is used to collect production data during the mold injection process, including the process parameters of the injection molding machine and the mold status data; the production data is synchronously mapped to the digital twin for dynamic simulation, and the deformation trend of the key areas of the mold is predicted online to obtain the trend prediction results;

[0013] The coordinated adjustment scheme matching unit is used to reverse analyze and trace the influencing factors and contributions that cause the predicted deformation based on the trend prediction results, and generate source analysis results; based on the trend prediction results and source analysis results, it dynamically matches coordinated adjustment schemes from a pre-set coordinated control strategy library.

[0014] The self-learning optimization execution unit is used to coordinate and adjust the plan, continuously collect production data after execution and feed it back to the digital twin, compare the actual deformation of the mold after adjustment with the predicted target, and determine the adjustment error; use the adjustment error to perform self-learning optimization on the simulation model of the digital twin and the adjustment strategy library.

[0015] Thirdly, the present invention also provides an electronic device including a processor and a memory, the memory being used to store computer program code, the computer program code including computer instructions, wherein when the processor executes the computer instructions, the electronic device performs the method as described in the first aspect above and any possible implementation thereof.

[0016] Fourthly, the present invention also provides a computer-readable storage medium storing a computer program, the computer program including program instructions that, when executed by a processor of an electronic device, cause the processor to perform a method as described in the first aspect above and any possible implementation thereof.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0018] 1) By constructing a virtual mold digital twin that includes multi-physics field coupling simulation capabilities of heat, fluid and force, and combining it with real-time process parameters and mold status data collected during the injection molding process, synchronous mapping and dynamic simulation of production data and digital twin are achieved. Compared with traditional single-physics field simulation models, this significantly improves the accuracy of predicting deformation trends in key areas of the mold. At the same time, the online prediction mode completely solves the prediction lag problem of traditional technology, improves the accuracy and real-time performance of mold deformation trend prediction, and provides accurate and timely data support for subsequent control.

[0019] 2) Based on the deformation trend prediction results, by reverse analysis to trace the deformation influencing factors and their contribution, the root cause of deformation can be accurately located. Then, the adjustment scheme is dynamically matched from the pre-set collaborative control strategy library. This changes the blindness of traditional single parameter adjustment and realizes multi-dimensional precise collaborative control of the root cause of deformation. It effectively improves the pertinence and effectiveness of deformation control, significantly reduces the deformation in key areas of the mold, and ensures the dimensional accuracy of injection molded products.

[0020] 3) By collecting and feeding back production data after regulation to the digital twin, the regulation error is determined by comparing the actual deformation with the predicted target. This allows for self-learning optimization of the simulation model and regulation strategy library, forming a closed-loop system of "data acquisition - simulation prediction - regulation execution - error feedback - model optimization." This system enables the simulation accuracy and regulation strategy adaptability of the digital twin to continuously improve with production conditions, gradually reducing regulation errors. Long-term operation can significantly improve the stability and adaptability of the production process, reduce product defect rates, and simultaneously reduce reliance on human experience, thereby increasing production efficiency.

[0021] 4) Precise deformation control through self-learning optimization can effectively avoid structural damage caused by excessive mold deformation, reduce mold maintenance frequency and wear, and extend mold life. At the same time, the improvement of product qualification rate and production efficiency, as well as the reduction of manual control costs, achieve optimized control of production costs from multiple dimensions and improve the economic benefits of enterprise production.

[0022] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the background art, the accompanying drawings used in the embodiments of the present invention or the background art will be described below.

[0024] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the specification, serve to illustrate the technical solutions of this disclosure.

[0025] Figure 1A flowchart illustrating an intelligent design method for precision injection molds based on the coordinated control of deformation fields, provided in an embodiment of the present invention;

[0026] Figure 2 This is a schematic diagram of the structure of a precision injection mold intelligent design system based on the coordinated control of deformation field, provided in an embodiment of the present invention. Detailed Implementation

[0027] To enable those skilled in the art to better understand the present invention, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0029] Please see Figure 1 , Figure 1 This is a flowchart illustrating an intelligent design method for precision injection molds based on the coordinated control of deformation fields, provided as an embodiment of the present invention. Figure 1 As shown, the method includes:

[0030] S10. Based on the geometric model, material properties, sensor layout, and historical production data of the mold, construct a virtual mold model that includes multi-physics field coupling simulation capabilities of heat, fluid, and force, as a digital twin;

[0031] S20. Collect production data during the mold injection process, including the process parameters of the injection molding machine and the mold status data; synchronously map the production data to the digital twin for dynamic simulation, predict the deformation trend of key areas of the mold online, and obtain the trend prediction results;

[0032] S30. Based on the trend prediction results, reverse analyze and trace the influencing factors and contributions that cause the prediction distortion, and generate source analysis results; based on the trend prediction results and source analysis results, dynamically match collaborative adjustment schemes from the pre-set collaborative control strategy library.

[0033] S40. Execute the collaborative adjustment plan, continuously collect production data after execution and feed it back to the digital twin, compare the actual deformation of the mold after adjustment with the predicted target, and determine the adjustment error; use the adjustment error to perform self-learning optimization on the simulation model of the digital twin and the adjustment strategy library.

[0034] This solution aims to adopt a technical approach of "digital twin construction → online prediction → traceability analysis → intelligent control → closed-loop optimization" to achieve real-time prediction, intelligent control, and continuous optimization of mold deformation during injection molding by constructing a high-precision virtual mold model.

[0035] In existing technologies, mold simulation models are typically based on idealized geometric models and material parameters, and mostly rely on static simulation models, failing to fully reflect the complex and ever-changing thermo-mechanical-fluid coupling effects in actual production. While conventional CAD / CAE software can perform basic injection molding process simulations, its model parameters are often based on idealized assumptions and lack a dynamic interaction mechanism with actual production data. Therefore, step S10 employs a virtual mold model based on the mold's geometric model, material properties, sensor layout, and historical production data, constructing a simulation capability encompassing thermo-fluid-mechanical multiphysics coupling. By calibrating the model parameters using historical production data, the digital twin accurately reflects the dynamic response characteristics of the actual mold, solving the problem of mismatch between traditional simulation models and the actual mold's dynamic response. This step first obtains the precise geometric structure of the mold from a 3D CAD model and converts it into a mesh model suitable for finite element analysis. Subsequently, based on key parameters such as the thermal conductivity, specific heat capacity, coefficient of thermal expansion, elastic modulus, and Poisson's ratio of the actual mold material, corresponding material property definitions are assigned to the mesh model to ensure the accuracy of physical properties. Building upon this foundation, and considering the characteristics of the mold cavity structure, gating system, cooling system, and ejection system, a multiphysics-coupled control equation for melt flow, heat transfer, and structural mechanics is established. This forms a computational framework capable of reflecting the real physical process. Compared to traditional mold design methods, the core of this step lies in directly integrating sensor placement into the virtual model. Temperature, pressure, and displacement monitoring points are set at key nodes of the finite element mesh, achieving a precise mapping between the physical and digital worlds. Furthermore, historical production data is used to dynamically calibrate the boundary conditions, initial conditions, and material parameters in the multiphysics-coupled control equation, enabling the virtual mold model to continuously approximate the actual dynamic response characteristics of the physical mold. This process solves the problem of the disconnect between the simulation model and the actual production environment in traditional mold design, laying a solid foundation for subsequent real-time prediction and control.

[0036] In traditional methods, production data is disconnected from the simulation model, making real-time online prediction impossible. Step S20 first uses a sensor network deployed at key locations on the mold to collect real-time injection molding machine process parameters (including injection speed, injection pressure, holding pressure, holding time, melt temperature, etc.) and mold status data (including temperature distribution, pressure changes, and micro-deformation data). This real-time data is transmitted to the digital twin via an industrial IoT interface, dynamically updating the boundary and initial conditions of the virtual model, achieving synchronized operation between the physical production system and the virtual simulation environment. Based on the updated digital twin, the thermo-fluid-mechanical multiphysics coupling control equations are solved using explicit or implicit time integration methods, achieving real-time dynamic simulation of the mold injection process. Through this process, the system can predict online the temperature, pressure, and displacement field distributions of key areas such as the mold cavity surface, parting surface, and ejection system within a specific future time step, identifying potential deformation deviation areas in advance. Compared to traditional offline simulation, this online prediction capability allows mold designers to monitor the mold's dynamic response in real time, providing a valuable time window for proactive intervention. Therefore, step S20, by collecting injection molding machine process parameters and mold status data in real time and simultaneously mapping them to a digital twin for dynamic simulation, can predict the deformation trend of key areas of the mold online. This step utilizes multiphysics coupling simulation technology to achieve deep integration of production data and simulation models.

[0037] Step S30 primarily provides the process of reverse tracing and matching control strategies. Existing mold adjustment technology relies heavily on manual experience and lacks systematic deformation tracing analysis methods. When product quality issues arise, technicians often need to go through multiple trial and error processes to find an effective solution, which is inefficient and costly. This solution, through data-driven reverse tracing analysis, can quickly and accurately locate the key factors causing deformation deviations, and, combined with a rich strategy library, generate targeted control solutions, significantly improving the efficiency and accuracy of problem solving.

[0038] Specifically, when the system identifies a potential deformation deviation in a certain area of ​​the mold, it initiates a multi-dimensional source analysis mechanism. This process first constructs a complete set of influencing factors, including injection molding process parameters (such as melt temperature, injection pressure, and holding pressure curve), mold structure parameters (such as cooling channel layout and cavity thickness), material parameters (such as material shrinkage rate and coefficient of thermal expansion), and environmental parameters (such as ambient temperature), among other possible factors. Then, a contribution quantification method based on sensitivity analysis is used to perform controlled variable simulations in a digital twin to calculate the contribution of each influencing factor to a specific deformation mode. Based on the source analysis results, the system dynamically matches the optimal adjustment scheme from a pre-set collaborative control strategy library. This collaborative control strategy library integrates historical successful cases, expert experience, and control strategies generated by multi-objective optimization algorithms. The matching process comprehensively considers the current deformation mode, major influencing factors, and their control priorities, generating a collaborative adjustment scheme that includes injection molding process parameter adjustments, mold temperature control schemes, and mold structure compensation schemes. This intelligent matching mechanism ensures the targetedness and effectiveness of the control scheme, avoiding the blindness of traditional trial-and-error adjustments.

[0039] The final step, S40, involves continuous optimization following the implementation of the collaborative adjustment scheme, forming a complete closed-loop self-learning mechanism—the key to the system's continuous evolution. During the collaborative adjustment scheme execution phase, the system sends control commands to the injection molding machine control system and the mold temperature control system, enabling injection molding production under the adjusted process conditions. Simultaneously, a sensor network continuously collects mold state data after adjustment, including temperature, pressure, and deformation data in key areas, and feeds this actual response data back to the digital twin. By comparing the actual deformation of the mold after adjustment with the predicted target, the system calculates the absolute error, relative error, and error distribution, forming a quantitative evaluation of the control effect. This evaluation result not only judges the effectiveness of the current control scheme but also provides a data foundation for the self-optimization of the digital twin. Based on the control error analysis, the system uses a data assimilation method to correct the parameters of the multiphysics coupling control equations of the virtual mold model, enabling the simulation model to continuously approximate the real characteristics of the physical mold. Simultaneously, reinforcement learning algorithms are used to continuously optimize the collaborative control strategy library, updating strategy parameters and matching rules. This dual self-learning mechanism ensures that the system can continuously adapt to changes in mold state and fluctuations in production conditions as operating time accumulates, exhibiting increasingly stronger prediction accuracy and control effects. Compared to static optimization systems, this dynamic self-learning capability enables intelligent mold design methods to continuously evolve, providing technical assurance for the long-term stable production of high-quality injection molded products.

[0040] Therefore, the solutions provided in the above embodiments construct a digital twin with multi-physics field coupling simulation capabilities including heat, fluid, and force, achieving high-precision deformation prediction; establish a real-time mapping mechanism between production data and simulation models, realizing online dynamic simulation; propose a deformation tracing method based on sensitivity analysis, which can quickly locate key influencing factors; design a collaborative control strategy library and a multi-objective optimization algorithm, realizing intelligent control; and construct a closed-loop self-learning optimization mechanism, enabling the system to continuously improve performance.

[0041] In one embodiment, the construction of a virtual mold model based on the mold's geometric model, material properties, sensor layout, and historical production data, which includes multiphysics-physics coupling simulation capabilities (thermal-fluid-mechanical), comprises:

[0042] Based on the 3D CAD model of the mold, a finite element mesh model of the mold is established;

[0043] Based on the thermal conductivity, specific heat capacity, coefficient of thermal expansion, elastic modulus and Poisson's ratio parameters of the mold material, the finite element mesh model is assigned corresponding material properties;

[0044] Based on the mold cavity structure, gating system, cooling system and ejection system, multi-physics field coupled control equations for melt flow, heat transfer and structural mechanics are established;

[0045] Based on the sensor layout, temperature, pressure and displacement monitoring points are set at the corresponding nodes of the finite element mesh model;

[0046] By using historical production data, the boundary conditions, initial conditions, and material parameters in the multiphysics coupling control equations are calibrated, and a virtual mold model consistent with the dynamic response of the physical mold is established.

[0047] In this embodiment, based on the 3D CAD model of the mold, the continuous geometric model is transformed into discrete mesh elements using finite element meshing technology. This process first requires importing the mold's geometric model, and then determining the appropriate mesh size and type based on simulation accuracy requirements and computational resources. For injection molds, more refined meshing is needed for cavity surfaces and structurally complex areas, while sparser meshing can be used for relatively simple structural areas to achieve a balance between computational efficiency and accuracy. Through finite element meshing, the complex mold geometry is transformed into a computational model composed of nodes and elements, providing a foundation for subsequent physics simulation. Reasonable meshing ensures accurate capture of deformation and temperature distribution in key areas of the mold, while maintaining controllable computational resources, thus laying the foundation for accurate simulation.

[0048] Furthermore, based on parameters such as thermal conductivity, specific heat capacity, coefficient of thermal expansion, elastic modulus, and Poisson's ratio of the mold material, corresponding material properties are assigned to the finite element mesh model. This step is crucial for establishing an accurate simulation model, as the thermophysical and mechanical properties of the material directly affect the mold's response behavior during injection molding. Specifically, thermal conductivity affects the mold's heat transfer efficiency and determines the cooling rate; specific heat capacity reflects the material's heat storage capacity; the coefficient of thermal expansion relates temperature changes to dimensional changes; and the elastic modulus and Poisson's ratio determine the mold's deformation characteristics under pressure. These parameters collectively constitute the material's multiphysics characteristics, forming the basis for simulation accuracy. Then, based on the mold cavity structure, gating system, cooling system, and ejection system, multiphysics coupling control equations for melt flow, heat transfer, and structural mechanics are established. Multiphysics coupling analysis overcomes the limitations of traditional single-field analysis, enabling more realistic simulation of complex physical phenomena during injection molding. For example, melt flow affects temperature distribution, temperature changes affect material properties and structural deformation, and structural deformation, in turn, affects the flow path, forming complex coupling relationships. By establishing coupled control equations, the virtual mold can simultaneously solve for the temperature field, flow field, and stress field, more accurately predicting the mold's behavior under actual working conditions and providing a reliable basis for deformation prediction and optimization.

[0049] Next, based on the sensor placement, temperature, pressure, and displacement monitoring points are set at the corresponding nodes of the finite element mesh model. This step establishes a data bridge between the physical mold and the virtual model, enabling the digital twin to remain synchronized with the real mold. The setting of monitoring points needs to consider the actual installation positions of the sensors on the physical mold, ensuring a one-to-one correspondence between the monitoring points in the virtual model and the measurement points on the physical mold. In this way, temperature, pressure, and displacement data collected from the physical mold can be directly mapped to the corresponding positions in the virtual model, achieving real-time data-driven simulation.

[0050] Finally, historical production data was used to calibrate the boundary conditions, initial conditions, and material parameters in the multiphysics coupling control equations, establishing a virtual mold model consistent with the dynamic response of the physical mold. This step is crucial for improving model accuracy, and the simulation model is continuously optimized through data-driven methods.

[0051] Historical production data contains information on the mold's response under actual working conditions. This information is used to correct parameters in the simulation model, such as boundary conditions (cooling water temperature, injection pressure, etc.), initial conditions (mold preheating temperature, etc.), and material parameters (considering performance variations under actual working conditions). Through repeated calibration, the virtual model gradually approximates the real behavior of the physical mold, reducing errors caused by idealized assumptions. This calibration method based on historical data enables the virtual mold model to continuously learn and optimize; as production data accumulates, the model's predictive accuracy continuously improves.

[0052] The above process constructs a high-precision virtual mold model. Through multiphysics coupling simulation and real-time data-driven methods, accurate prediction of the mold's working state is achieved. Compared with traditional design methods, this approach significantly reduces the number of trial runs, improves mold design efficiency, and lowers production costs.

[0053] In one embodiment, the step of synchronously mapping production data to a digital twin for dynamic simulation and online prediction of deformation trends in key areas of the mold includes:

[0054] Real-time acquisition of injection molding machine process parameters, including injection speed, injection pressure, holding pressure, holding time, and melt temperature; real-time acquisition of mold status data, including temperature sensor data, pressure sensor data, and displacement sensor data for key areas;

[0055] The collected production data is synchronously mapped to the corresponding boundary conditions and initial conditions of the virtual mold model; among which,

[0056] Based on the multiphysics field coupled control equations, the implicit time integration method is used for dynamic simulation calculations;

[0057] It outputs the temperature, pressure and displacement field distributions of key areas such as the mold cavity surface, parting surface and ejection system in real time, and predicts the deformation trend in the future time step.

[0058] First, the system collects two main types of data in real time through a high-precision sensor network deployed in key areas of the mold (such as the cavity surface and near the parting line):

[0059] Injection molding machine process parameters include injection speed, injection pressure, holding pressure, holding time, and melt temperature. These are the core variables driving the injection molding process.

[0060] Mold status data: The physical state changes of key areas of the mold are directly monitored through temperature, pressure and displacement sensors.

[0061] Furthermore, the collected real-time data is synchronously mapped into the virtual mold model, serving as the boundary and initial conditions for its dynamic simulation. That is, parameters such as injection pressure and melt temperature in the virtual model are completely consistent with the parameters actually set on the injection molding machine. Local temperature, pressure, and deformation data of the mold measured by sensors are used to calibrate and drive the state of corresponding nodes in the virtual model. Through this mapping, the digital twin is no longer a static, idealized model, but a dynamic mirror reflecting the real working conditions of the physical mold in real time. This significantly improves the initial accuracy and realism of the simulation model.

[0062] The system is based on pre-established thermo-fluid-mechanical multiphysics coupled control equations and employs an implicit time integration method for dynamic simulation calculations. Implicit methods are better suited for handling the inherent rigidity and nonlinearity issues in injection molding, offering better computational stability and allowing for larger time steps, thus improving computational efficiency while maintaining accuracy. Multiphysics coupled simulation can accurately describe the interactions between melt flow, heat transfer, and structural deformation. For example, it can simulate the impact of melt shear heat generation on the mold temperature field, and how thermal stress caused by temperature changes leads to mold deformation.

[0063] Based on high-fidelity dynamic simulation, the system can output real-time temperature, pressure, and displacement distribution cloud maps of key areas such as the mold cavity surface, parting surface, and ejection system. More importantly, it can predict the deformation trend of the mold over several time steps based on the current physical state and by solving the governing equations. This elevates mold condition monitoring from passive perception to proactive prediction and early warning. For example, the system can predict in advance that heat accumulation at a certain point on the parting surface may cause warping exceeding tolerances, providing a valuable time window for proactive intervention. Through the closed-loop linkage of the above four steps, this embodiment integrates real-time perception, high-fidelity mapping, accurate simulation, and forward-looking prediction. Its ultimate effect is to achieve a paradigm shift from "monitoring deformations that have occurred" to "predicting deformations that will occur," providing accurate and timely decision-making basis for subsequent intelligent control and compensation, thereby effectively reducing the number of trial moldings and improving product quality and control efficiency.

[0064] In one embodiment, the step of reverse analysis and tracing back the influencing factors and their contributions that lead to the prediction distortion based on the trend prediction results includes:

[0065] Based on the trend prediction results, identify the key areas and deformation modes of mold deformation;

[0066] Based on key regions and deformation modes, a set of influencing factors is constructed, including injection molding process parameters, mold structure parameters, material parameters, and environmental parameters;

[0067] Using the set of influencing factors as input, sensitivity analysis is employed to perform controlled variable simulations in a digital twin to calculate the contribution of each influencing factor to the deformation of the key area.

[0068] Based on the contribution ranking, the main influencing factors that cause the predicted deformation and their influence paths are identified, and the source analysis results are output.

[0069] First, based on trend prediction results obtained from digital twin simulations, such as displacement field cloud maps of the mold cavity surface and parting surface, the system automatically identifies key areas where deformation exceeds a preset threshold. Then, by analyzing the displacement vector distribution and curvature changes in these areas, it accurately determines whether the deformation mode is warping, shrinkage, or local collapse. This step defines the specific targets and scope for subsequent source tracing analysis. The system then constructs a comprehensive set of multi-dimensional influencing factors around the identified key areas and deformation modes. This set includes not only real-time adjustable injection molding process parameters (such as melt temperature, injection pressure, and holding pressure curves), but also relatively fixed mold structure parameters (such as cooling channel layout and cavity thickness), material parameters (such as material shrinkage rate and coefficient of thermal expansion), and environmental parameters (such as workshop ambient temperature). This systematic construction method ensures the comprehensiveness of the source tracing analysis and avoids tracing bias caused by missing factors.

[0070] The system takes the aforementioned set of influencing factors as input and performs controlled variable simulations in a digital twin. Specifically, while keeping other factors constant, a certain influencing factor is adjusted at a certain step size (e.g., increasing the melt temperature by 5°C), and the changes in deformation in the key area are observed and recorded. Subsequently, a sensitivity coefficient is used to quantify the contribution of each factor. Specifically, the contribution is calculated as follows:

[0071] ;

[0072] The change in deformation caused by a change in a certain influencing factor is obtained through simulation;

[0073] : Baseline deformation, the deformation value under initial simulation conditions;

[0074] The amount of change in this influencing factor;

[0075] The baseline value of this impact factor.

[0076] This formula allows for the precise calculation of the percentage change in deformation caused by a unit percentage change in each influencing factor, thus enabling an objective and quantitative comparison of contributions. Based on contribution ranking, the system automatically identifies the main influencing factors whose cumulative contribution exceeds a set threshold (e.g., 70%). More importantly, the digital twin can trace the complete impact path of these main factors leading to deformation based on multi-physics coupled data. For example, it can trace the complete chain of "increased melt temperature → increased local thermal load on the mold → uneven thermal stress distribution → warping of the cavity surface." Ultimately, the system generates a structured source analysis report. This report clearly lists the main influencing factors causing the predicted deformation, their respective contributions, specific impact path descriptions, and may provide preliminary control recommendations based on historical data.

[0077] Compared to traditional experience-dependent mold problem-solving methods, this embodiment implements a data-driven, mechanism-clear, and quantitatively precise intelligent traceability analysis process. This significantly shortens the mold debugging cycle and greatly reduces over-reliance on specific expert experience, thereby enabling more precise control over mold production and improving productivity and stability.

[0078] In one embodiment, the step of dynamically matching a coordinated adjustment scheme from a pre-set coordinated control strategy library based on trend prediction results and source analysis results includes:

[0079] Based on the trend prediction results, determine the control targets for mold deformation, including the deformation control range and deformation mode optimization targets; based on the source analysis results, determine the set of controllable influencing factors and their control priorities.

[0080] Retrieve control strategies from the coordinated control strategy library that match the control objectives, influencing factor set, and control priority; optimize the parameters of the retrieved control strategies based on a multi-objective optimization algorithm to generate a coordinated adjustment scheme;

[0081] The coordinated adjustment scheme includes an injection molding process parameter adjustment scheme, a mold temperature control scheme, and a mold structure compensation scheme.

[0082] In this embodiment, the system first identifies key deformation areas requiring intervention (such as cavity surfaces and parting surfaces) and their specific deformation modes (such as warpage and shrinkage) based on trend prediction results provided by the digital twin. Based on this, and considering the dimensional tolerances and quality requirements of the injection molded product, the system quantifies and determines the specific objectives to be achieved in this adjustment, such as controlling the deformation in a specific area within ±0.1 mm or eliminating asymmetrical warpage modes. Simultaneously, the system sets boundary constraints for the adjustment. These constraints typically stem from the performance limits of the injection molding machine (such as maximum injection pressure and screw speed range) and the safety thresholds of the mold itself (such as the highest temperature and maximum stress the material can withstand), ensuring that the generated adjustment plan is executed within a safe and feasible range.

[0083] Subsequently, based on the information provided by the source analysis results—namely, the ranking of the contributions of each influencing factor to the current deformation—the system filters out parameters that can be adjusted in real-time or near real-time on the actual production line, forming a "subset of adjustable influencing factors." The system then determines the adjustment priority of these adjustable factors. Priority ranking considers not only the magnitude of contribution but also a comprehensive evaluation of the adjustment response speed and cost of each factor. Next, the system uses the adjustment objectives, constraints, adjustable factor subsets, and adjustment priorities determined in the first two steps as composite query conditions to search and match within a pre-set collaborative adjustment strategy library. This strategy library integrates standardized adjustment strategies constructed from historical successful cases, expert experience, and validated simulation data. The matching process aims to find one or more basic adjustment strategies that best match the current deformation pattern and adjustable factor combination, providing an initial solution for subsequent refined optimization. This step changes the traditional trial-and-error approach relying on individual engineer experience, realizing intelligent strategy recommendation based on a large amount of historical data and knowledge, thus improving the scientific nature and efficiency of decision-making.

[0084] For the matched basic strategy, the system uses a multi-objective optimization algorithm to optimize the parameters and generate the final coordinated adjustment scheme. The optimization process constructs an evaluation function that includes multiple objectives, such as deformation control accuracy, production cycle efficiency, and system energy consumption. Under set constraints, the optimization algorithm automatically searches for the strategy parameter combination that achieves the optimal balance among these objectives. The final generated coordinated adjustment scheme is a comprehensive solution, typically explicitly including:

[0085] Injection molding process parameter adjustment schemes: such as adjusting injection speed, holding pressure curve, etc.

[0086] Mold temperature control scheme: such as adjusting the flow rate or temperature setpoint of each cooling circuit.

[0087] Mold structure compensation scheme: When necessary, provide micro-compensation schemes that can be executed online, such as local compensation for cavity deformation by adjusting the opening sequence of valves in specific cooling circuits.

[0088] Thus, the above embodiments can achieve a leap from "single-objective, experience-based decision-making" to "multi-objective, automatic optimization," and the generated solutions can simultaneously take into account quality, efficiency, and cost, reduce over-reliance on the experience of senior engineers, and effectively improve the consistency of product quality and the stability of the production process.

[0089] In a preferred embodiment, when generating the coordinated adjustment scheme based on the trend prediction results and the source analysis results, a deformation sensitivity compensation factor is also introduced to dynamically correct the adjustment amount of the process parameters, specifically including:

[0090] For the main influencing factors identified in the source analysis, such as holding pressure, melt temperature, or cooling water temperature, calculate their local sensitivity to the maximum predicted deformation in the key area:

[0091] ;

[0092] : Represents the maximum deformation response caused by a unit change in process parameters, which is estimated online through digital twin differential simulation or finite difference method;

[0093] : The maximum displacement of the key region predicted by the digital twin in the current cycle;

[0094] : No. Adjustable process parameters, such as holding pressure.

[0095] Then, based on this sensitivity, a deformation sensitivity compensation factor is defined. :

[0096] ;

[0097] : , is the empirical scaling factor.

[0098] Multiply the initial adjustment recommendation by The actual adjustment amount is obtained:

[0099] ;

[0100] The final execution command sent to the injection molding machine or temperature control system is the actual adjustment amount executed.

[0101] Adjustment amount initially generated by strategy library matching or optimization algorithms.

[0102] Therefore, this embodiment can automatically suppress overshoot in the high-sensitivity region and allow a larger adjustment range in the low-sensitivity region, avoiding oscillations or overshoot caused by linear control.

[0103] See Figure 2 In one embodiment, the present invention also provides an intelligent design system for precision injection molds based on the coordinated control of deformation fields, the system comprising:

[0104] The digital twin building unit 100 is used to construct a virtual mold model with thermal-fluid-force multiphysics coupling simulation capabilities based on the mold's geometric model, material properties, sensor layout, and historical production data, serving as a digital twin.

[0105] The deformation trend prediction unit 200 is used to collect production data during the mold injection process, including the process parameters of the injection molding machine and the mold status data; the production data is synchronously mapped to the digital twin for dynamic simulation, and the deformation trend of the key areas of the mold is predicted online to obtain the trend prediction results;

[0106] The coordinated adjustment scheme matching unit 300 is used to reverse analyze and trace the influencing factors and contributions that cause the predicted deformation based on the trend prediction results, and generate source analysis results; based on the trend prediction results and source analysis results, it dynamically matches coordinated adjustment schemes from a pre-set coordinated control strategy library.

[0107] The self-learning optimization execution unit 400 is used to coordinate and adjust the scheme, continuously collect production data after execution and feed it back to the digital twin, compare the actual deformation of the mold after adjustment with the predicted target, and determine the adjustment error; use the adjustment error to perform self-learning optimization on the simulation model of the digital twin and the adjustment strategy library.

[0108] It is understood that the system provided in this embodiment has functions or includes modules that can be used to execute the methods described in the above method embodiments. The specific implementation can be referred to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0109] The present invention also provides an electronic device including a processor and a memory, the memory being used to store computer program code, the computer program code including computer instructions, wherein when the processor executes the computer instructions, the electronic device performs a method as described in any of the above possible implementations.

[0110] The present invention also provides a computer-readable storage medium storing a computer program, the computer program including program instructions that, when executed by a processor of an electronic device, cause the processor to perform a method as described in any of the above possible implementations.

[0111] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

Claims

1. A method for intelligent design of precision injection molds based on synergistic control of deformation field, characterized in that, The method includes: Based on the geometric model, material properties, sensor layout, and historical production data of the mold, a virtual mold model with multi-physics coupling simulation capabilities including heat-fluid-force is constructed as a digital twin. Collect production data during the mold injection process, including the process parameters of the injection molding machine and the mold status data; synchronously map the production data to a digital twin for dynamic simulation, predict the deformation trend of key areas of the mold online, and obtain the trend prediction results; Based on the trend prediction results, the influencing factors and their contributions that cause the prediction distortion are analyzed in reverse and traced back to generate source analysis results; based on the trend prediction results and source analysis results, a collaborative adjustment scheme is dynamically matched from a pre-set collaborative control strategy library; The collaborative adjustment plan is implemented, and production data after implementation is continuously collected and fed back to the digital twin. The actual deformation of the mold after adjustment is compared with the predicted target to determine the adjustment error. The adjustment error is used to perform self-learning optimization of the simulation model of the digital twin and the adjustment strategy library.

2. The intelligent design method for precision injection molds based on synergistic control of deformation field according to claim 1, characterized in that, The virtual mold model, which incorporates thermo-fluid-mechanical multiphysics coupling simulation capabilities, is constructed based on the mold's geometric model, material properties, sensor layout, and historical production data. This includes: Based on the 3D CAD model of the mold, a finite element mesh model of the mold is established; Based on the thermal conductivity, specific heat capacity, coefficient of thermal expansion, elastic modulus and Poisson's ratio parameters of the mold material, the finite element mesh model is assigned corresponding material properties; Based on the mold cavity structure, gating system, cooling system and ejection system, multi-physics field coupled control equations for melt flow, heat transfer and structural mechanics are established; Based on the sensor layout, temperature, pressure and displacement monitoring points are set at the corresponding nodes of the finite element mesh model; By using historical production data, the boundary conditions, initial conditions, and material parameters in the multiphysics coupling control equations are calibrated, and a virtual mold model consistent with the dynamic response of the physical mold is established.

3. The intelligent design method for precision injection molds based on synergistic control of deformation field according to claim 1, characterized in that, The process of synchronously mapping production data to a digital twin for dynamic simulation and online prediction of deformation trends in key areas of the mold includes: Real-time acquisition of injection molding machine process parameters, including injection speed, injection pressure, holding pressure, holding time, and melt temperature; real-time acquisition of mold status data, including temperature sensor data, pressure sensor data, and displacement sensor data for key areas; The collected production data is synchronously mapped to the corresponding boundary conditions and initial conditions of the virtual mold model; among which, Based on the multiphysics field coupled control equations, the implicit time integration method is used for dynamic simulation calculations; It outputs the temperature, pressure and displacement field distributions of key areas such as the mold cavity surface, parting surface and ejection system in real time, and predicts the deformation trend in the future time step.

4. The intelligent design method for precision injection molds based on synergistic control of deformation field according to claim 1, characterized in that, The process of reverse-engineering and tracing the influencing factors and their contributions that lead to prediction distortions based on trend prediction results includes: Based on the trend prediction results, identify the key areas and deformation modes of mold deformation; Based on key regions and deformation modes, a set of influencing factors is constructed, including injection molding process parameters, mold structure parameters, material parameters, and environmental parameters; Using the set of influencing factors as input, sensitivity analysis is employed to perform controlled variable simulations in a digital twin to calculate the contribution of each influencing factor to the deformation of the key area. Based on the contribution ranking, the main influencing factors that cause the predicted deformation and their influence paths are identified, and the source analysis results are output.

5. The intelligent design method for precision injection molds based on synergistic control of deformation field according to claim 1, characterized in that, The dynamic matching of coordinated adjustment schemes from a pre-set coordinated control strategy library based on trend prediction results and source analysis results includes: Based on the trend prediction results, determine the control targets for mold deformation, including the deformation control range and deformation mode optimization targets; based on the source analysis results, determine the set of controllable influencing factors and their control priorities. Retrieve control strategies from the coordinated control strategy library that match the control objectives, influencing factor set, and control priority; optimize the parameters of the retrieved control strategies based on a multi-objective optimization algorithm to generate a coordinated adjustment scheme; The coordinated adjustment scheme includes an injection molding process parameter adjustment scheme, a mold temperature control scheme, and a mold structure compensation scheme.

6. A precision injection mold intelligent design system based on deformation field collaborative control, characterized in that, The system includes: The digital twin building block is used to construct a virtual mold model with multi-physics coupling simulation capabilities, including thermal-fluid-mechanical fields, based on the mold's geometric model, material properties, sensor layout, and historical production data, thus serving as a digital twin. The deformation trend prediction unit is used to collect production data during the mold injection process, including the process parameters of the injection molding machine and the mold status data; the production data is synchronously mapped to the digital twin for dynamic simulation, and the deformation trend of the key areas of the mold is predicted online to obtain the trend prediction results; The coordinated adjustment scheme matching unit is used to reverse analyze and trace the influencing factors and contributions that cause the predicted deformation based on the trend prediction results, and generate source analysis results; based on the trend prediction results and source analysis results, it dynamically matches coordinated adjustment schemes from a pre-set coordinated control strategy library. The self-learning optimization execution unit is used to coordinate and adjust the plan, continuously collect production data after execution and feed it back to the digital twin, compare the actual deformation of the mold after adjustment with the predicted target, and determine the adjustment error; use the adjustment error to perform self-learning optimization on the simulation model of the digital twin and the adjustment strategy library.

7. An electronic device, characterized in that, include: A processor and a memory, the memory being used to store computer program code, the computer program code including computer instructions, wherein when the processor executes the computer instructions, the electronic device executes the intelligent design method for precision injection molds based on the coordinated control of deformation fields as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which includes program instructions that, when executed by a processor of an electronic device, cause the processor to perform the intelligent design method for precision injection molds based on the collaborative control of deformation fields as described in any one of claims 1 to 5.

Images